68% of enterprise AI fails here—here's how to fix it
summary
Claims that 68% of enterprise AI stalls on fragmented data and shows a live build that avoids complex MLOps, connecting real-time Salesforce Notion Zendesk Confluence and docs to a semantic pipeline, producing grounded answers with citations and handling a MongoDB error and a delayed product question. The demo constructs an embeddable AppSmith agent embedded in Zendesk, creates Jira tickets, notifies Slack, and reports a shipped sales assistant that cuts response times by 72% faster response time.
AI-written summary · about
chapters (8)
- 0:00Introduction
- 1:46Barriers to AI adoption
- 3:11Meeting enterprises with AI
- 3:32Live RAG pipeline demo
- 6:46Adding data sources and agents
- 7:48Embedding AI in workflows
- 9:39AppSmith agents and privacy
- 9:59Closing remarks
related talks
50mRealities & Misconceptions about Enterprise Applications of AI/ML and Data
28mWhy AI Works for Individuals but Fails in Enterprises, and How to Fix It.
27mBuilding AI applications: Successes and Lessons
21mOvercoming Data Barriers for Successful AI Projects
20mThe Three Ways AI Agents Fail. Here's How to Fix Yours.
26mAccelerating DevOps with AI
18mFrom POC to Profit: Deploying Customer-Facing AI Agents That Actually Work
39mFrom Demo to Production: Why Agentic AI Systems Fail - and How to Fix Them